In a color recognition system, it has been conventional to irradiate a test surface with light from a light source, for example as shown in U.S. Pat. No. 4,464,054 to Karras et al., which is incorporated herein by reference. The reflected light from the test surface is directed to a plurality of photo detectors through differently colored filters. In general, three colors of filters, customarily red, green and blue, are used. These three colors make up the color components of the sample. The electrical output from the photo detectors vary with respect to the intensity of each color component in the reflected light and provide a quantitative measurement for each of the color components and therefore a quantitative measurement of color for the test surface.
It is common in the art to calibrate a color recognition system by using a standard tile, or standard test surface. The quantitative value for each color component of the standard tile is known and programmed into the color recognition system. The system is calibrated by initially sampling the standard tile and comparing the sampled color component values to the known standard color component values. Differences between two sets of data establish specific correction factors for each color component. Calibration of future tests with respect to the standard tile are accomplished by using these correction factors.
This technique assumes that the light source illuminates each test surface with light having identical color component characteristics as the light used during calibration. In other words, this technique assumes the light source will flash consistently for each test. However, there may be some variation of the light source from sample to sample. When the light source varies from the standard characteristics used during calibration, the test values must not only be calibrated with respect to the standard tile, but also should be adjusted to compensate for the variation of the light source.